期刊追踪 | MIS Q: Management Information Systems: 2023年第47卷4期

算法厌恶是一个重要而持久的问题，它阻止了从人工智能的进步中获益。迄今为止，文献提供的解释主要集中在有意识的反思过程上。在这里，我们通过一个无意识的视角来补充这个观点，这个视角可以提供很多信息。基于隐性偏见的理论，在一项预先注册的研究中，我们认为人们对人工智能(AI)系统产生了一种隐性偏见(即偏见)，将其视为一种不同的、具有威胁性的“物种”，其行为是未知的。就像在其他偏见的环境中一样，我们期望人们被这种隐含的偏见所引导，但试图超越它。这导致了一些依赖算法建议(欣赏)的意愿，这种意愿随着人们对机器的隐性偏见而减少。接下来，基于躯体标记假说和可及性-诊断性观点，我们提供了厌恶为何短暂的解释。随着人们了解算法的性能，他们在决定是否依赖人工智能的建议时，不再那么依赖原始的隐含偏见。两项使用内隐联想测验的研究(n1 = 675，n2 = 317)一致支持这一观点。另外两项研究(n3 = 255，n4 = 332)排除了替代解释，为我们的主张提供了更有力的支持。这些发现最终表明，在厌恶和欣赏之间移动指针最初取决于一个人对人工智能的普遍无意识偏见，因为没有足够的信息来推翻它。他们进一步指出，在后期的使用阶段，这种转变取决于对人工智能性能诊断信息的可访问性，这减少了无意识偏见的权重。

Algorithm aversion is an important and persistent issue that prevents harvesting the benefits of advancements in artificial intelligence. The literature thus far has provided explanations that primarily focus on conscious reflective processes. Here, we supplement this view by taking an unconscious perspective that can be highly informative. Building on theories of implicit prejudice, in a preregistered study, we suggest that people develop an implicit bias (i.e., prejudice) against artificial intelligence (AI) systems, as a different and threatening “species,” the behavior of which is unknown. Like in other contexts of prejudice, we expected people to be guided by this implicit bias but try to override it. This leads to some willingness to rely on algorithmic advice (appreciation), which is reduced as a function of people’s implicit prejudice against the machine. Next, building on the somatic marker hypothesis and the accessibility-diagnosticity perspective, we provide an explanation as to why aversion is ephemeral. As people learn about the performance of an algorithm, they depend less on primal implicit biases when deciding whether to rely on the AI’s advice. Two studies (n1 = 675, n2 = 317) that use the implicit association test consistently support this view. Two additional studies (n3 = 255, n4 = 332) rule out alternative explanations and provide stronger support for our assertions. The findings ultimately suggest that moving the needle between aversion and appreciation depends initially on one’s general unconscious bias against AI because there is insufficient information to override it. They further suggest that in later use stages, this shift depends on accessibility to diagnostic information about the AI’s performance, which reduces the weight given to unconscious prejudice.

本文旨在解决人类-人工智能(AI)增强中迄今为止被忽视的一个方面:机器诱导的反射。通过建立一个基于理论的机器诱导反射模型，我们对信息系统中正在进行的关于人工智能的讨论和反射理论的研究做出了贡献。在我们的多阶段研究中，医生使用基于机器学习的(ML)临床决策支持系统(CDSS)来观察这种交互是否以及如何在X射线诊断任务的背景下刺激反思实践。通过分析口头协议、绩效指标和调查数据，我们开发了一个综合的理论基础来解释基于ML的系统如何帮助刺激反思实践。个人根据他们是否察觉到与这些CDSS系统的冲突或一致，参与更批判或更浅薄的模式，这反过来导致不同层次的反思深度。通过揭示机器诱导思考的过程，我们为研究这种基于人工智能的系统如何帮助个人变得更具反思性，从而更有效地成为专业人士提供了一个不同的视角。这一观点与传统的基于人工智能的决策支持系统的效率观形成鲜明对比，也丰富了人工智能增强的理论。

This paper addresses a thus-far neglected dimension in human-artificial intelligence (AI) augmentation: machine-induced reflections. By establishing a grounded theoretical-informed model of machine-induced reflection, we contribute to the ongoing discussion in information systems (IS) regarding AI and research on reflection theories. In our multistage study, physicians used a machine learning-based (ML) clinical decision support system (CDSS) to see if and how this interaction can stimulate reflective practice in the context of an X-ray diagnosis task. By analyzing verbal protocols, performance metrics, and survey data, we developed an integrative theoretical foundation to explain how ML-based systems can help stimulate reflective practice. Individuals engage in more critical or shallower modes depending on whether they perceive a conflict or agreement with these CDSS systems, which in turn leads to different levels of reflection depth. By uncovering the process of machine-induced reflections, we offer IS research a different perspective on how such AI-based systems can help individuals become more reflective, and consequently more effective, professionals. This perspective stands in stark contrast to the traditional, efficiency-focused view of ML-based decision support systems and also enriches theories on human-AI augmentation.

 在美国，多医院系统(MHSs)对医院服务的收费明显高于独立医院。竞争限制理论认为，具有多市场联系的厂商可以暗中勾结，相互抑制价格竞争，以保持价格高于竞争水平。我们假设，这种MMC诱导的竞争约束(休战)的成功受到it在公司层面和市场单位层面的两个冲突角色的影响。公司母公司寻求在企业范围内标准化IT应用程序，以协调市场单位，作为联合实施竞争限制策略并在企业范围内保持高价格的一种手段。然而，市场单位，即聚集在地理患者市场的MHS成员医院，面临着降低服务成本的竞争压力。市场单位寻求使用差异化的IT应用程序来实现成本降低，这进而加剧了当地市场的价格竞争，危及休战的可持续性，并削弱了公司母公司的竞争约束战略对整个企业的价格影响。在2005-2013年期间对美国195个多医院系统的纵向研究中，我们发现了对这些观点的支持。MHS操作IT的全公司标准化补充了竞争限制策略，以提高全公司的价格。市场单位对差异化分析信息技术的使用降低了当地市场的成本，削弱了竞争限制策略的价格效应。该研究通过理论化多单位、多市场(MUMM)组织的企业级和市场单位级it如何对MMC和MUMM组织收取的平均价格之间的联系产生相反的调节作用，推进了研究和实践。

In the U.S., multihospital systems (MHSs) charge significantly higher prices for hospital services than stand-alone hospitals. Rivalry restraint theory suggests that MHS with multimarket contact (MMC) can tacitly collude and mutually forebear from price competition to keep their prices above competitive levels. We posit that the success of such MMC-induced rivalry restraints (the truce) is affected by two conflicting roles of IT at the corporate level and market unit levels, respectively. The corporate parent seeks to standardize IT applications enterprise-wide to coordinate market units as a means of jointly implementing the rivalry restraint strategy and keeping prices high enterprise-wide. However, market units, i.e., the member hospitals of MHS clustered in geographic patient markets, face competitive pressures to reduce their service costs. Market units seek to use differentiated IT applications to achieve cost reductions, which then fuel price competition in local markets, jeopardize the sustainability of the truce, and weaken the enterprise-wide price effects of the corporate parent’s rivalry restraint strategy. In a longitudinal study of 195 multihospital systems in the U.S. in the 2005-2013 time period, we found support for these ideas. The corporate-wide standardization of the operational IT of MHS complements the rivalry restraint strategy to increase enterprise-wide prices. Market units’ use of differentiated analytical IT reduces costs in local markets and weakens the price effects of the rivalry restraint strategy. The study advances IS research and practice by theorizing how the corporate-level and the market unit-level IT of a multi-unit, multimarket (MUMM) organization can have opposing moderating effects on the link between MMC and the average prices charged by the MUMM organization.

 随着金融科技的快速发展，动态信用风险评估的需求变得越来越重要。虽然以前的信用评分研究主要集中在单期贷款违约预测上，但我们呼吁采用一种新的方法——多期违约预测(MPDP)——来描述一段时间内的风险状况。为了解决MPDP提出的挑战，如单调违约概率预测和复杂关系调节，我们提出了一种新的方法，混合和集体评分(HACS)。我们设计了一种混合建模策略，分别通过违约判别模型和违约时间估计模型来预测借款人是否会违约以及何时违约，并通过一个概率框架来综合它们。为了适应各种可能的违约时间模式和度量连续时间间隔上的违约概率分布，我们提出了一种联合违约建模方法来训练违约时间估计模型。模型(违约时间预测性能和判别性能)和机制(可识别性和可判别性)级别的实证评估，以及应用(授予性能和盈利性能)级别的影响分析表明，HACS在所有方面都优于基准生存分析和多标签学习方法。它可以更准确地预测违约时间，并为金融机构和投资者在发放贷款和选择贷款组合时提供更好的决策支持。

With the rapid development of fintech, the need for dynamic credit risk evaluation is becoming increasingly important. While previous studies on credit scoring have mostly focused on single-period loan default prediction, we call for a new avenue—multiperiod default prediction (MPDP)—to depict risk profiles over time. To address the challenges raised by MPDP, such as monotonic default probability prediction and complex relationship accommodation, we propose a novel approach, hybrid and collective scoring (HACS). We design a hybrid modeling strategy to predict whether and when a borrower will default separately through a default discrimination model and a default time estimation model, respectively, and synthesize them through a probabilistic framework. To accommodate various possible patterns of default time and measure the distribution of default probability over successive time intervals, we propose a joint default modeling method to train the default time estimation model. Empirical evaluations at the model (time-to-default prediction performance and discrimination performance) and mechanism (identifiability and discriminability) levels, as well as impact analyses at the application (granting performance and profitability performance) level, show that HACS outperforms the benchmarked survival analysis and multilabel learning methods on all fronts. It can more accurately predict time-to-default and provide financial institutions and investors better decision-support in granting loans and selecting loan portfolios.

 我们利用企业基于注意力的观点来检验首席信息官在高管团队中的存在是否以及何时对企业的构思数字创新(IDI)(即企业数字专利活动的强度)和商业化数字创新(CDI)(即企业新产品的数字复杂性)产生积极影响。基于注意力过程依赖于情境的观点，我们还探索了CEO特征(IT背景和角色任期)以及环境特征(行业IT注意力)的调节作用。我们分析了美国标准普尔500公司跨行业小组八年来的数据，其中包括多达2，852家公司的年度观察。结果表明，首席信息官在高层管理团队中的存在与企业的IDI和CDI正相关。此外，他们表明，与首席执行官特征相关的组织背景调节首席信息官-CDI的关系，而与行业IT关注相关的环境背景调节首席信息官-IDI的关系。我们的研究为信息系统文献做出了贡献，提供了有力的证据证明首席信息官在高管团队中的存在对企业的数字创新成果有积极影响，展示了内部和外部边界条件如何影响首席信息官的工作，并阐述了管理注意力作为解释数字创新的潜在机制的作用。

We draw on the attention-based view of the firm to examine whether and when the presence of a CIO in the TMT has a positive effect on both firms’ ideated digital innovation (IDI) (i.e., the intensity of firms’ digital patenting activity) and commercialized digital innovation (CDI) (i.e., the digital sophistication of firms’ new products). Building on the idea that attention processes are context dependent, we also explore the moderating roles of CEO characteristics (IT background and role tenure) as well as environmental characteristics (the industry’s IT attention). We analyze data from a cross-industry panel of U.S. S&P 500 firms over eight years that includes up to 2,852 firm-year observations. The results indicate that CIO presence in the TMT is positively related to a firm’s IDI and CDI. Furthermore, they show that the organizational context related to CEO characteristics moderates the CIO-CDI relationship and that the environmental context related to the industry’s IT attention moderates the CIO-IDI relationship. Our research contributes to the information systems literature by providing robust evidence that CIO presence in the TMT positively influences a firm’s digital innovation outcomes, showing how internal and external boundary conditions affect the work of CIOs, and elaborating the role of managerial attention as an underlying mechanism explaining digital innovation.

 追求金融包容性对于加快经济增长和增强金融可持续性的重要性已得到充分注意。然而，对于金融机构如何平衡盈利性和平等性之间的潜在社会经济权衡，研究几乎没有提供可操作的见解。一个主要挑战是，除了经济盈利能力之外，缺乏对各种市场信息对财务平等的影响的了解。另一个挑战在于如何评估现实世界中大量反事实政策下的社会经济权衡。我们本次研究的动机是源于移动设备和互联网接入的高度渗透的数字化用户行为数据(即“替代数据”)的新兴来源。因此，我们调查了来自智能手机和社交媒体的替代数据如何帮助减轻潜在的金融不平等，同时在金融信用风险评估的背景下保持企业盈利能力。我们与一家领先的小额贷款网站合作，设计了一个新颖的“元”实验，允许我们在一组详尽的反事实政策下模拟各种真实世界的实地实验。有趣的是，我们发现使用智能手机活动描述用户金融风险在提高金融包容性方面比使用在线社交媒体信息有效1.3倍(好23.05%比18.11%)，在提高业务盈利能力方面有效近1.3倍(好42%比33%)。令人惊讶的是，我们发现使用消费者的在线购物活动进行信用风险分析会损害金融包容性。此外，我们调查了金融包容性改善的潜在解释。我们的发现表明，替代数据，尤其是用户的智能手机活动，不仅表现出更高的普遍性，而且似乎更正交于传统的敏感人口统计属性。这反过来有助于减轻在基于机器的风险评估过程中由未观察到的因素或代表性不足的训练样本驱动的统计偏差。

The importance of pursuing financial inclusion to accelerate economic growth and enhance financial sustainability has been well noted. However, studies have provided few actionable insights into how financial institutions can balance the potential socioeconomic trade-off between profitability and equality. One major challenge arises from a lack of understanding of the impacts of various types of market information available on financial equality beyond economic profitability. Another challenge lies in how the socioeconomic trade-off under a large set of counterfactual policies in a real-world setting can be evaluated. Our motivation for the present study was the emerging sources of digitized user-behavior data (i.e., “alternative data”) stemming from the high penetration of mobile devices and internet access. Accordingly, we investigated how alternative data from smartphones and social media can help mitigate potential financial inequality while preserving business profitability in the context of financial credit risk assessment. We partnered with a leading microloan website to design a novel “meta” experiment that allowed us to simulate various real-world field experiments under an exhaustive set of counterfactual policies. Interestingly, we found that profiling user financial risk using smartphone activities is 1.3 times more effective in improving financial inclusion than using online social media information (23.05% better vs. 18.11%), and nearly 1.3 times more effective in improving business profitability (42% better vs. 33%). Surprisingly, we found that using consumers’ online shopping activities for credit risk profiling can hurt financial inclusion. Furthermore, we investigated potential explanations for financial inclusion improvements. Our findings suggest that alternative data, especially users’ smartphone activities, not only demonstrate higher ubiquity but also appear to be more orthogonal to conventional sensitive demographic attributes. This, in turn, can help mitigate statistical bias driven by the unobserved factors or underrepresentative training samples in machine-based risk assessment processes.

 尽管由于先进信息技术的使用，多渠道销售策略已经变得很普遍，但是一种交易机制如何影响另一种交易机制的结果，尤其是在B2B市场，仍然没有得到充分的探索。本文研究了来自在线标价预售渠道的价格和数量信息对已有百年历史的荷兰顺序拍卖系统性能的影响。卖家可以控制支付的价格，并在拍卖预售中提供一定比例的股票。预售后剩下的任何东西都通过拍卖出售。我们对近150万件花卉拍品的分析揭示了一个积极的影响，即预售拍品的拍卖价格和总收入高于非预售拍品。这一结果甚至适用于在预售中没有实际销售的批次，表明买家密切关注来自标价预售渠道的额外信息。通过梳理预售价格和预售数量对拍卖价格的信息影响，我们评估了一些定价策略。结果表明，预售时高价销售仍然比打折销售更有利。

Although multichannel sales strategies have become common due to the use of advanced information technologies, how one trading mechanism can influence the outcome of another, especially in the B2B market, remains largely underexplored. This paper investigates the effect of price and quantity information from an online posted-price presales channel on the performance of the century-old sequential Dutch auction system. Sellers can control the price paid and make a proportion of their stock available in auction presales. Anything left after presales is sold via auctions. Our analysis of nearly 1.5 million flower lots reveals a positive effect with higher auction prices and total revenue for lots listed in presales than for lots that are not. The result holds even for lots with no actual sales in the presales, indicating that buyers pay close attention to the additional information from the posted-price presales channel. By teasing out the information effect of presales prices and presales quantity on auction prices, we evaluate a number of pricing strategies. The results suggest that selling at a high price in presales is still more beneficial than selling more by discounting prices.

 根据信息隐私、发展心理学和家庭科学的文献，本研究调查了父母的在线隐私担忧如何传递给青少年，并影响他们在社交网站上的自我表露。我们认为，父母的隐私担忧直接(即顺从)和通过青少年的隐私担忧(即内化)间接减少青少年的自我表露，这种影响受到亲子隐私不和谐、父母互联网评估中介和青少年性别的调节。为了测试研究模型，我们从中国726个家庭收集了匹配的亲子数据。研究结果表明，父母的隐私关注通过影响青少年的隐私关注间接影响青少年的自我表露。此外，亲子隐私不和谐削弱了青少年隐私担忧对自我表露的影响。父母使用互联网评估中介来指导青少年在线活动的程度加强了父母隐私关注对青少年隐私关注的影响。统计分析进一步表明，青少年隐私担忧的中介效应被亲子隐私不和谐削弱，但被互联网评价性中介加强。我们还发现，父母的隐私担忧通过不同的途径影响儿子和女儿，尤其是当父母使用高互联网评估中介时。在高互联网评价中介下，父母隐私关切主要通过间接路径(通过儿子的隐私关切)影响儿子的自我表露，但通过女儿的隐私关切直接或间接影响女儿的自我表露。最后，我们讨论了理论贡献和实践意义。

Drawing upon the literatures in information privacy, developmental psychology, and family science, this research investigates how parental online privacy concerns can be passed on to adolescents and affect their self-disclosure on social networking sites. We propose that parental privacy concerns decrease adolescents’ self-disclosure both directly (i.e., compliance) and indirectly through adolescents’ privacy concerns (i.e., internalization) and that such effects are moderated by parent-child privacy dissonance, parental internet evaluative mediation, and adolescents’ gender. To test the research model, we collected matched parent-child data from 726 families in China. The results show the indirect effect of parental privacy concerns on adolescents’ self-disclosure via their influence on adolescents’ privacy concerns. In addition, parent-child privacy dissonance weakens the effect of adolescents’ privacy concerns on self-disclosure. The extent to which parents employ internet evaluative mediation to guide adolescents’ online activities reinforces the effect of parental privacy concerns on adolescents’ privacy concerns. Statistical analyses further revealed that the mediating effect of adolescents’ privacy concerns is weakened by parent-child privacy dissonance but strengthened by internet evaluative mediation. We also found that parental privacy concerns affect sons and daughters through different paths, especially when parents employ high internet evaluative mediation. Under high internet evaluative mediation, parental privacy concerns affect sons’ self-disclosure primarily through an indirect path (via sons’ privacy concerns), but influence daughters’ self-disclosure both directly and indirectly via daughters’ privacy concerns. We conclude by discussing theoretical contributions and practical implications.

 心流是人们在全身心投入地行动时所体验到的整体感觉，它是像任务绩效这样的预期工作结果的已知驱动力。然而，IT在工作中的日益普及可能会扰乱员工的流动。因此，信息技术介导的中断对心流的影响值得在研究和实践中给予更多的关注。我们进行了一项NeuroIS实验室实验，专注于一项典型的办公室工作任务——发票匹配任务(即，将客户付款与发票匹配)。我们操纵中断频率(低、高)和内容相关性(不相关、相关)来研究中断对自我报告血流、其维度和高频心率变异性的影响(HF-HRV；从心电图记录计算)作为副交感神经系统(PNS)激活的代表。我们发现内容相关性调节了中断频率和自我报告流之间的关系，并且这些结果随着流维度而变化。内容相关性也调节了中断频率和PNS激活之间的关系。此外，自我报告的心流与感知的和客观的任务绩效正相关，而PNS激活与任何一种绩效测量都无关。最后，我们发现PNS激活(通过高频HRV测量)和自我报告的血流之间没有关系，这有助于神经病学文献中关于生理学证据是否构成自我报告的替代或补充的重要辩论。总的来说，我们的发现表明，频繁的打扰本身是无害的。相反，考虑内容相关性对于更全面地理解对自我报告流的影响、其维度和潜在生理学是至关重要的。

Flow, the holistic sensation people experience when they act with total involvement, is a known driver for desired work outcomes like task performance. However, the increasing ubiquity of IT at work can disrupt employees’ flow. Thus, the impact of IT-mediated interruptions on flow warrants more attention in research and practice. We conducted a NeuroIS laboratory experiment focusing on a typical office work task—an invoice matching task (i.e., matching customer payments to invoices). We manipulated interruption frequency (low, high) and content relevance (irrelevant, relevant) to study the impact of interruptions on self-reported flow, its dimensions, and high-frequency heart rate variability (HF-HRV; calculated from electrocardiography recordings) as a proxy for parasympathetic nervous system (PNS) activation. We found that content relevance moderated the relationship between interruption frequency and self-reported flow and that these results vary along flow dimensions. Content relevance also moderated the relationship between interruption frequency and PNS activation. Furthermore, self-reported flow was positively associated with both perceived and objective task performance, while PNS activation was not related to either performance measure. Lastly, we found no relationship between PNS activation (measured by HF-HRV) and self-reported flow, contributing to an important debate in the NeuroIS literature on whether physiological evidence constitutes an alternative or a complement to self-reports. Overall, our findings indicate that frequent interruptions are not harmful per se. Rather, considering content relevance is critical for a more comprehensive understanding of the effects on self-reported flow, its dimensions, and the underlying physiology.